Comments (17)
Graph.norm()为0表明兼容图中没有边,所有匹配互不兼容。匹配的采样数不要设置太小
from 3d-registration-with-maximal-cliques.
1、采样数量小一些,比如2500
2、兼容阈值大一点,比如0.999
from 3d-registration-with-maximal-cliques.
hello, 目前测试修改后,降采样为0.7,采样5000个数,兼容性得分0.999,不过配准效果还是不太好,请问可能是什么原因~?地面点、噪声影响?
from 3d-registration-with-maximal-cliques.
你的数据有gt吗?建议和真值比较看下误差。MAC是粗配准方法,如果要进行高精度配准,还需要icp进一步优化结果
from 3d-registration-with-maximal-cliques.
@zhangxy0517 好的,谢谢您的建议。还想请问一下在日记中,哪一个参数是用来评估配准效果好坏?,或者论文中,类似得分之类的信息?
raw_correspondence:3468
correspondence:250
Start registration.
进入第二个registration
进入第二个Graph_construction
graph construction: 0.0216231
coefficient computation: 0.000375635
0.0908374->min(0.0908374 0.82656 0.510234)
clique computation: 7.6468e-05
clique selection: 0.000262457
hypothesis generation & evaluation: 0.000587079
14 : 14
3 2.00638
0.224797 0.50723 0.986243
-0.985204 -0.0913466 -0.145011 -3.2895
0.109196 -0.986711 -0.120321 2.15954
-0.132094 -0.134375 0.982087 0.656984
0 0 0 1
from 3d-registration-with-maximal-cliques.
RE和TE,RMSE能评估误差。目前demo函数没有与gt评估的功能。代码中提供了评估函数:
calculate_rotation_error
calculate_translation_error
RMSE_compute_scene
如需使用可以参考实验代码
from 3d-registration-with-maximal-cliques.
@zhangxy0517 好的,感谢您的回复!
from 3d-registration-with-maximal-cliques.
hello,请问一下,代码中有提供评价 两帧匹配收敛效果好坏的评价指标吗, 是根据 double score = evaluation_trans() 这个函数 判断吗,
from 3d-registration-with-maximal-cliques.
对的,evaluation_trans()根据匹配来评估每个位姿变换假设,分数最高的作为最终结果
from 3d-registration-with-maximal-cliques.
好的,那最终是看迭代后的分best_score 来评估吗,如果A和多帧进行配准,得分都是best_score=1 ,那么如何挑选配准最佳的?
from 3d-registration-with-maximal-cliques.
MAC是两两配准方法。evaluation_trans评估的是一对点云产生的所有假设的分数,并不适用于多对点云
from 3d-registration-with-maximal-cliques.
不可以的
from 3d-registration-with-maximal-cliques.
可以参考下python版本的feature matching
from 3d-registration-with-maximal-cliques.
@zhangxy0517 您好,feature matching c++版本的搜索KD tree这部分,我看其时间复杂度是n*nlogn, python版本看似加快了运算, 但是两个描述子在做内积运算的时候,其时间复杂度是n^3, 然后在找出每行最大, python版本的时间复杂度好像更慢呀?
from 3d-registration-with-maximal-cliques.
算法对点云的数据有要求吗?比如激光的数据和预测深度的数据(没有激光数据好),包括稀疏程度等 都适用吗?谢谢
from 3d-registration-with-maximal-cliques.
MAC的输入是特征匹配,理论上对于不同数据类型是通用的。但是也要具体问题具体分析,建议多尝试。
from 3d-registration-with-maximal-cliques.
好的,谢谢
from 3d-registration-with-maximal-cliques.
Related Issues (20)
- error report in python version HOT 9
- How to run this code?
- Linux version compiling error HOT 14
- RE < 15 and TE < 30? HOT 1
- About U3M dataset HOT 4
- Out of memory when testing 3DMatch Dataset with FPFH descriptors. HOT 4
- 数据集链接似乎失效了 HOT 2
- 请求提供更多的实验结果
- 请求提供geotransformer处理的3dmatch/3dlomatch数据集处理结果
- Parameter Conversion Errors: when Running C++ Code with igraph, PCL on Windows HOT 1
- 关于版本MSVC =2022 x64
- 您好,关于demo运行时间问题 HOT 1
- python调用c++版本mac出现segmentation falut
- 你好,关于在Linux中编译时igraph中出现了类型错误 HOT 4
- Problems with reproduction of registration results FPFH+MAC from the paper HOT 2
- 作者您好,我在结合FCGF和MAC在3DMatch数据集上测试了一下发现将体素大小设置为0.05的情况,部分最大团数量过多(2000万以上),这情况不知道您是否遇到,然后我将体素大小设为0.08,测试完和性能远不如原论文的性能,作者您当时是怎么解决这个问题的 HOT 7
- Wrong registration results. What can I do? HOT 3
- Graph Construction部分十分耗时 HOT 2
- 运行MAC的方式
- 请问python版本的和linux版本的性能一致吗? HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from 3d-registration-with-maximal-cliques.